An Improved Approach for Vehicles Detection and Tracking

This paper presents an approach for moving vehicles detection and tracking in outdoor scenes. Firstly the projection curves of the frame difference image of two consecutive frames was analyzed to find the approximate areas of all moving vehicles and then the template matching with matching pel count (MPC) was used in target tracking. Finally, verification system ensures tracking correctly. If necessary, the approach segments all moving targets. The proposed approach, in experiments of tracking vehicles in traffic scenes, shows good efficiency and precision. It detects moving vehicles quickly and tracks them exactly